Why does cygwin and linux give different array outputs?

carlr · 08-08-2012, 12:43 PM

Hi,

I have written a script which works in cygwin (CYGWIN_NT-6.1-WOW64 1.7.11(0.260/5/3)) but i know want to run it on lunix (Linux 2.6.18-53.1.21.el5 #1 SMP Tue May 20 09:35:07 EDT 2008).

The script which works in cygwin does not work in linux. After some investigation I have found it to be due to array issues, so i wrote a simpel test script:

#!/bin/bash

PART_NAME=Wheel.inp

eval ARRAY_2D_ELSET=( $(awk -F, '{print "["$1"]=\""$0"\" "}' "$PART_NAME"_2D_ELSET.tmp) )

for ((i=1; i<=10; i++)) do
echo ${ARRAY_2D_ELSET[i]} >> ARRAYout.txt
done

The output files vary when run on linux and cygwin. This is causing my script not to function.

Linux output:
1]= 1, 183, 3, 4
[
2]= 2, 59, 357, 46
[
3]= 3, 91, 356, 170
[
4]= 4, 356, 39, 61
[
5]= 5, 94, 355, 163
[

Cygwin output:
1, 183, 3, 4
2, 59, 357, 46
3, 91, 356, 170
4, 356, 39, 61
5, 94, 355, 163
6, 355, 161, 206
7, 123, 354, 122
8, 354, 127, 144
9, 353, 129, 212
10, 104, 353, 24

Can any one explain to me why this script does not give the same output in both? Also, and more importantly, is there any way to make the linux output the same as the cygwin one? or do i have to go and make some significant edits to my origional cygwin script.

A sample of the file "$PART_NAME"_2D_ELSET.tmp is below:

1, 183, 3, 4
2, 59, 357, 46
3, 91, 356, 170
4, 356, 39, 61
5, 94, 355, 163
6, 355, 161, 206
7, 123, 354, 122
8, 354, 127, 144
9, 353, 129, 212
10, 104, 353, 24

dwhitney67 · 08-08-2012, 01:50 PM

I'm not a scripting expert, however I did take the script and data that you posted to see if I could recreate the problem you have described. The output generated under both Ubuntu 10.04 and Kubuntu 12.04 were the same as you described for cygwin.

ntubski · 08-08-2012, 02:23 PM

Is it possible some or all of your input files have DOS style (\r\n) line endings? What bash version are you using on Cygwin and Linux?

carlr · 08-08-2012, 03:15 PM

Linux:
GNU bash, version 3.1.17(1)-release (x86_64-redhat-linux-gnu)
Copyright (C) 2005 Free Software Foundation, Inc.

cygwin:
GNU bash, version 4.1.10(4)-release (i686-pc-cygwin)
Copyright (C) 2009 Free Software Foundation, Inc.

I've just checked and i have \n line endings on all files.

Kenhelm · 08-08-2012, 04:38 PM

I can reproduce the problem in Linux by putting a space at the beginning of each line of data.
Without the spaces I get the Cygwin output.

Code:

echo ' 1, 183, 3, 4
 2, 59, 357, 46
 3, 91, 356, 170
 4, 356, 39, 61
 5, 94, 355, 163
 6, 355, 161, 206
 7, 123, 354, 122
 8, 354, 127, 144
 9, 353, 129, 212
 10, 104, 353, 24' > file.tmp

Code:

eval ARRAY_2D_ELSET=( $(awk -F, '{print "["$1"]=\""$0"\" "}' file.tmp) )

for ((i=1; i<=10; i++)); do
  echo ${ARRAY_2D_ELSET[i]}
done

1]= 1, 183, 3, 4
[
2]= 2, 59, 357, 46
[
3]= 3, 91, 356, 170
[
4]= 4, 356, 39, 61
[
5]= 5, 94, 355, 163
[

carlr · 08-08-2012, 04:47 PM

Thank you for your help every one.

As Kenhelm says, it seems to be that cygwin can deal with a space in front to the first character, where as linux can not. i'm not sure why. It may just be the versions of bash i'm using.

devnull10 · 08-08-2012, 05:31 PM

A bit off topic but any reason why you're loading into an array and then looping through to print rather than just redirecting the output from awk?